📰On The Record – 10th August 2026

An unconfirmed leak out of China suggested Apple could raise iPhone 17 prices as soon as today, as the global memory chip crunch finally catches up with the world's most valuable phone. Apple has already pushed prices up on Macs and iPads this year while holding the line on iPhones — until now, apparently. The bigger story isn't one device getting pricier. It's that the AI boom's appetite for memory chips has grown so large it's now reaching directly into your pocket, and the iPhone 18 may not escape it either.

Moody's issued a blunt warning yesterday: banks racing to bolt AI onto everything from fraud detection to customer service are quietly making themselves dependent on a tiny cluster of tech and cloud providers. The upside is real — lower costs, higher revenues — but so is the downside, including outages, cybersecurity gaps and a 20% chance AI could be doing a mid-level banker's job by 2030. It's a reminder that "AI adoption" isn't just a software upgrade. It's banks quietly outsourcing a piece of their own resilience to companies they don't control.

Job-hopping has quietly ground to a halt this year, and it's not because workers suddenly fell back in love with their employers. It's fear. With hiring slower and AI reshaping entire departments, staying put has become the safer bet, even for people who'd rather be anywhere else. That's good news for retention numbers and terrible news for wage growth, since job-switching has always been the fastest way to a pay rise. Businesses are quietly benefiting from a workforce too nervous to leave, which might look like stability on a spreadsheet but rarely builds loyalty that lasts once the job market thaws.

The quiet part: a workforce staying out of fear isn't the same as one staying out of faith.

OpenAI halted parts of development on its upcoming Astra model after internal testing suggested it was nearing the company's highest-ever "critical" cybersecurity risk threshold — the kind of capability that could hunt down and exploit software vulnerabilities with barely any human help. The company says it's tightening safeguards before pushing further. It's a rare moment of a frontier AI lab hitting its own brakes rather than a regulator's, and it says something bigger: the industry is starting to bump up against risks it built the guardrails for, but never really expected to test.

Mondays are when good intentions either take shape or quietly die by 11am. This prompt is built to catch them before that happens — a quick weekly-planning exercise designed to turn a chaotic inbox and a vague sense of "lots to do" into three things that will actually move the needle. Copy it in, run it before your first meeting, and start the week already ahead of it.

Paste it in, fill in your priorities, and let this week start on your terms rather than your inbox's.

Change a single line of code and unlock 1,000+ AI models, smart routing for cost and performance, and persistent agent memory.

Book a Demo

Iran laid out tough conditions yesterday for reopening the Strait of Hormuz, demanding an end to U.S. military pressure, troop withdrawals, compensation and sanctions relief. It threatens to derail a fragile Iran-Oman deal aimed at restoring commercial shipping through one of the world's most critical oil corridors. For any business touched by energy costs or shipping timelines, this standoff is one to watch closely — Hormuz carries a huge share of the world's oil traffic.

A new Reuters/Ipsos poll found 61% of Americans back stronger government oversight of social media companies, with real bipartisan agreement behind it. Two-thirds support mandatory age verification to keep under-16s off platforms. The numbers reflect a growing consensus that's increasingly hard for lawmakers to ignore, and platforms should expect the regulatory pressure many have dodged for years to start landing with more force.

Israeli AI-security firm Irregular has become an unexpected common thread in recent security incidents touching OpenAI, Anthropic and Meta's models, after misconfigured testing environments let AI systems access the real internet and probe outside vulnerabilities. It's an uncomfortable reminder that the infrastructure built to safely stress-test powerful AI can itself become the weak point, and that "testing" and "real-world exposure" are getting harder to keep separate.

It's an inflation-heavy week, and the numbers on Wednesday and Thursday will likely set the tone for markets right through Friday.

Book more client interviews without adding researchers or coordinators. PodPitch finds relevant shows, personalizes outreach from your team’s inbox, and follows up until interviews land. Give clients more visible momentum, stronger authority, and another reason to renew.

Appear on 3,853,234 Podcasts

The world kept moving fast outside the boardroom too — from Middle East diplomacy to summer disaster season in full swing.

Around this time last year, President Trump publicly demanded Intel CEO Lip-Bu Tan resign over alleged ties to Chinese semiconductor firms, wiping value off the stock within hours of the post. Within a week Trump had reversed course entirely, meeting Tan at the White House and calling his career an "amazing story." It's a useful reminder heading into this year's chip supply crunch: in this industry, the ground can shift from crisis to calm in the time it takes to schedule a meeting.

The war for AI talent has stopped being a Silicon Valley story and become a full-blown economic one, with compensation packages for top researchers now rivalling those of professional athletes. It's easy to roll your eyes at eye-watering pay, but it reflects something real: there simply aren't enough people who can build these systems, and every company betting its future on AI knows it. The risk is that this scarcity distorts far more than a handful of salaries — it shapes which companies can compete at all, and which get priced out of the race entirely before it's even properly begun.

The talent war isn't a side story to the AI boom. Increasingly, it is the AI boom.

Bloomberg's latest breakdown of the memory chip crunch shows just how lopsided the AI buildout has become: hyperscalers are on track to spend roughly $650 billion in 2026 alone chasing high-bandwidth memory, and every gigabyte reserved for a data center is a gigabyte not going into your next phone or laptop. The chart makes the mechanism painfully clear — this isn't a temporary shortage, it's a structural reallocation of global chip supply toward AI infrastructure. Watch DRAM contract prices over the next quarter; if they keep climbing, expect more device makers to follow Apple's lead.